Exclusive Fujitsu has signed a deal to resell Violin Memory 6000-series networked flash arrays.
El Reg storage desk understands from a well-placed source that Fujitsu has a tripartite strategy for solid-state storage:
- It will use flash drives and caching in its Eternus storage arrays in a very fast data vault.
- Fusion-io PCIe flash storage hardware, together with Fusion's ioTurbine and directCache software, will be used to turbo-charge server applications.
Violin Memory 6000 arrays will accelerate server access to shared primary data, which needs access latencies lower than network disk-drive arrays can deliver, even ones with solid-state disks.
The directCache software is used to cache data from RAID arrays on ioMemory cards, ensuring faster read accesses. The ioTurbine software provides flash storage, also using ioMemory cards, for virtual machines in a server and supports VMware's vMotion.
Fujitsu, we understand, reckons its technology is similar to, but better than, EMC's VFCache product.
The Japanese company talked to various flash array suppliers, and thinks that, just as RAID boxes need their own custom hardware and management for the best performance, flash arrays need specific hardware and software to provide the best performance, reliability and longevity. This is why it picked Violin because, we were told, it offered "the best approach."
The Violin 3000 series was Violin's first flash array product and it wasn't really enterprise class; the later 6000 product line featured redundant controllers, as well as Symantec storage and data management software, making it an enterprise-class product.
It's our understanding that Garry Veale was involved with the Fujitsu-Violin deal as Violin's top EMEA executive - but overall, our contact said, "it was a huge team effort". Fujitsu has never seen such a deal negotiated and signed so quickly.
This agreement follows closely behind HP deciding not to extend a Violin Memory 3000 reselling deal beyond August 2014 nor expand it by adding in Violin's newer and faster 6000 products.
Casting a little more light on the HP-Violin relationship, Fujitsu learned that the Hewlett-Packard deal was always limited to the 3000 product and there was no discussion of HP reselling the later 6000. We understand that, because of this, there is no need for Violin to amend any filed papers for its rumoured IPO; some industry watchers had suggested these documents pointed to future lucrative revenues from HP, which will instead concentrate on its 3PAR arrays.
In fact, as the Fujitsu-Violin deal was only signed, we understand, yesterday; it represents a new and incremental revenue stream for Violin and that itself might necessitate a reworking of any IPO paperwork.
There are proof-of-concept products underway with some selected Fujitsu customers in specific but unknown verticals. The company was confident the deal would be done quite early and has already started necessary work to prepare for general availability of the Violin 6000 arrays.
We also understand that Fujitsu sees two key areas for the array's use: database and virtual desktop acceleration. Running specific applications in Violin's arrays, apps that need extremely low latency, is seen as a valid use case but not one that Fujitsu is focussed on.
The Fujitsu-Violin reselling deal is likely to be announced next week. It might well be worth going to SNW in Frankfurt, on 29 to 30 October, and checking out the Fujitsu stand. ®